In a communications network, information (e.g., voice or data) in the form of packets is routed between endpoints through one or more switching devices (e.g., routers and telecommunications switches). To prevent unauthorized access to the communications network, network managers have begun implementing Intrusion Detection systems. Intrusion Detection systems (IDS) continually monitor traffic in the communications network and identify known patterns of attack utilizing various detection techniques. When an intrusion is detected, the IDS typically notifies a network administrator and implements other predetermined or programmed responses. For example, the IDS may generate an alarm, report the attack, log the attack and/or terminate the intruder's connection.
The IDS is typically implemented on a server attached to the communications network. The server may be any type of computer on which intrusion detection software is loaded. The intrusion detection software examines packets entering and exiting the communications network to detect any anomalies, such as traffic load, TCP port usage or matching “signatures” of known types of intrusion attacks. If an anomaly is detected, the incoming or outgoing packet flow is labeled inappropriate. As used herein, the term packet flow refers to a unidirectional stream of packets sent between a source endpoint and a destination endpoint.
However, implementing the IDS on a server separate from the switching devices in the communications network limits the IDS processing capability and performance. For example, when positioning the IDS at the edge of the communications network, the communications links between the IDS and the external network and between the IDS and the switching devices each have limited bandwidth. In addition, the IDS server itself has limited processing capability. Therefore, as the traffic on the communications network increases, the limitations of bandwidth and processing capability either prevents the growth of the communications network or necessitates the use of additional IDS servers to perform intrusion detection for the communications network.
Recently, IDS boards have become available that can be directly implemented on the network switches and routers. The IDS boards provide the same functionality as the IDS server, but may be applicable to only a single switch or router. Therefore, aggregate data between multiple switching devices of the type normally used to identify intrusion attacks is more difficult to accumulate, which may result in missed intrusion alerts. In addition, the IDS boards require a physical space on the switching devices that could be used for other applications. The additional hardware also adds cost and complexity to the switching devices.